Efficacy to Implementation in the Child Welfare System

1
Efficacy to Implementation in the Child Welfare
System

• J. Landsverk, P. Chamberlain, J. Reid
• (Eve Reider, Chair)
• PSMG
• October 26, 2005
• Child and Adolescent Services Research Center,
  San Diego, CA
• Oregon Social Learning Center Center for
  Research to Practice, Eugene, OR
• National Institute on Drug Abuse, Rockville,
  MD

2

• Cascading Dissemination of a Foster Parent
  Intervention (R01 MH 60195)
• Principal Investigator Patti Chamberlain Co-Inve
  stigators Joe Price, John Reid, John Lansverk
• Funding Agency National Institute of Mental
  Health
• Preventing Health-Risking Behavior in Delinquent
  Girls (R01 DA15208-1)
• Principal Investigator Patti Chamberlain Co-Inve
  stigators John Reid, Leslie Leve
• Funding Agency National Institute of Drug Abuse
• Preventing Problems for Girls in Foster Care (R01
  MH54257-6)
• Principal Investigator Patti Chamberlain Co-Inve
  stigators Leslie Leve, John Reid
• Funding Agency National Institute of Mental
  Health
• Oregon Prevention Research Center (2 P30 MH
  46690)
• Principal Investigator John Reid Co-Investigator
  s Chamberlain, DeGarmo, Eddy, Fisher, Leve,
  Marinez, Fetrow, Patterson, Forgatch, Bank,
  Shortt, Capaldi
• Funding Agency National Institute of Mental
  Health
• Pathways Home Reducing Risk in the Child Welfare
  System (1 P20 DA017592)
• Principal Investigator John Reid Co-Investigator
  s Chamberlain, DeGarmo, Eddy, Fisher, Leve,
  Martinez, Fetrow, Bronz, Sprengelmeyer, and Smith
• Funding Agency National Institute of Mental
  Health
• Patterns of Youth Mental Health Care in Public
  Service Systems (U01 MH55282)
• Principal Investigator Richard Hough
• Funding Agency National Institute of Mental
  Health
• Mental Health Services Across Child Welfare
  Agencies (R01 MH59672)

3
Child Welfare Service Sector Characteristics
Related to Design Issues
4
Child Welfare as a Mission Context

• Effectiveness and implementation studies are
  service sector context specific (usual care),
  efficacy studies usually are not.
• Mission of Child Welfare
• Child Safety
• Child Permanence
• Child Well-Being
• Safety permanence services provided by CW
• Well-Being services provided by other sectors
  (mental health, substance abuse services)

5
Service Sector Comparisons on Risk level
Intervention Type, Reason for Entry Sex Ratio
6
Service Sector Comparison on Age at Entry
Most common entry age into service sector
7
Context of Child Welfare

• Yearly 5,000,000 referred, 3,000,000
  investigated (4.5 of all children), 826,000
  substantiated for child abuse and neglect
• Multiple living environments for child when
  services delivered (1) home of origin, (2)
  out-of-home care or foster care
  relative/non-relative, congregate care, (3)
  adoptive home
• Service trajectories may involve movement from
  one child setting to another

8
Nationally Representative Child Welfare Cohort
Study (NSCAW)

• 92 primary sampling units, 6,000 investigated
  cases (substantiated or indicated), followed for
  36 months
• 5 months after investigation
• 1 in residential care (congregate) (5.7 after
  36 months)
• 10 in relative or non-relative family foster
  care
• 24 received services in parents home
• 65 no further services beyond investigation

9
Need and Opportunity for Preventive Interventions

• High rates of mental disorders (42 for ages 6-17
  POC San Diego study, highest rates for ADHD,
  ODD, CD), developmental problems , social and
  cognitive problems
• At any given time, 1 of children under 18 live
  in out-of-home care, estimated to be 3-7 of all
  children at some time
• Externalizing problems, mission elements of
  safety and permanence gtgtgt parent mediated
  interventions

10
Child Welfare Context and Effectiveness Study
Design and Measurement Issues

• Multiple child settings multiple environmental
  contexts and multiple parent informants for
  longitudinal designs
• Clustering at entry multiple children removed
  from same family and placed in out-of-home care
  with confounding by maltreatment type ( more
  clustering for neglect)
• Clustering for case worker functions
  supervisory units of 8-10 case workers
• Timing for identification, sampling and enrolling
  because of early exit bias

11
Child Welfare Context and Effectiveness Study
Design Issues

• Little tradition of randomized study designs and
  use of standardized measurement
• Labor force (1) case workers - BA and MSW in
  social work, (2) para-professional foster parent
  (relative non-relative), congregate care shift
  workers (BA level)
• Cost analysis complicated by multiple sector
  costs
• safety and permanence costs are paid by child
  welfare
• parent training costs paid by education sector
  (community college)
• ameliorative costs (mental health, substance use,
  developmental) paid by other service sectors

12
Child Welfare Context and Effectiveness Study
Design Organization of Services

• Multiple levels state and county agencies,
  programs within agency, supervisory units within
  programs, families within case worker, children
  within families adjustment for clustering and
  nested designs
• Child nested within foster care and within
  biological family
• Service programs organized by type of child
  setting importance of change of setting and
  its meaning
• Entry into services
• Entry into out-of-home care
• Change of placement within out-of-home care
• Exit from out-of-home care
• Exit from services

13
Collaborations among Disparate Research Traditions

• Intervention Development and Efficacy Trials
  relatively small n studies, thick measures with
  triangulation
• Services Research and Clinical Epidemiology
  large N studies, thin measures including
  administrative data
• Prevention Science Methodologists

14
Developmental modeling and efficacy trials
leading to effectiveness and Implementation
trials of an intervention for foster parents in
the CWS
15
(No Transcript)
16
Early trials of PMT

• Crafted on basis of a great deal of clinical and
  systematic observation in homes
• Case studies and small randomized trials.
• Very thick measurement

17
Replications Across Developmental Level and Sites

• OSLC first graders through adolescents
• Wahler children
• Webster-Stratton preschoolers
• Eyeberg
• Forehand

18
Thick, Multi-Method Methodology

• Naturalistic and laboratory observations
• Parent, child, and teacher reports
• Administrative data from schools and juvenile
  courts
• Psychiatric data
• Analogue Tasks

19
Word Slide of Neglect Model
.38
.71
-.55
.36
.18
.56
log IPC
e12
Mother to
Child
.08
-.28
.47
.79
.30
.41
.20
.27
.13
-.33
.36
-.15
.62
.68
.66
.20
CBCL T
e13b
D4b
.75
.25
Child
.27
e14b
Scenarios
Child
Aggression
.20
WB
Interviewer
e15b
Rating
cmin5356.00,df\df,p\p,cmindf\cmindf,cfi\cfi
.39
r
2
1.29
TRF T
e17b
20
Mediated model for lift, playground etc
21
(No Transcript)
22
(No Transcript)
23
Academic Progress
24
CD AP DSM IV
Criteria

25
Foster Parents as Agents of Change

• Multi-dimensional treatment Foster Care

26
Multidimensional Treatment Foster Care (MTFC)

• For youth placed in out-of-home care
• Youth are placed singly in intensively trained
  and supervised community foster homes that are
  contacted daily and supported 24/7 for 69 mo.
• Interventions are implemented using multiple
  methods (e.g., family and individual therapy,
  skill training, academic supports) in key
  settings
• Program supervisors carry a caseload of 10,
  supervise foster parents, therapists, skills
  trainers, and work with parole/probation officers
• Youth attend public schools

27
MTFC Effects for Boys

• More time in program/fewer runaways
• Less time in locked incarceration in follow-up
• Fewer criminal offenses (½ the rate of GC boys)
• Less likely to commit violent crimes 2 years
  later
• Delinquency effects mediated by
• Supervision
• Relationship with a mentoring adult
• Consistent non-harsh discipline
• Less association with delinquent peers
• Chamberlain Reid, 1998 Eddy Chamberlain,
  2000
• Eddy, Whaley, Chamberlain, 2004

28
MTFC Effects for Girls

• Delinquency
• Deviant Peer Association
• School Attendance Homework Time
• Leve, Chamberlain Reid (in press), J of
  Counseling and Clinical Psychology Leve
  Chamberlain (2005) J of Abnormal Child
  Psychology
• Chamberlain Leve, in preparation

29
24-month arrest outcomes
6
5
5.2
4.5
4
Mean number
of arrests
3
2
1.4
1
1.3
0.8
0.7
0
GC
MTFC
Intervention group
12 months pre-treatment entry
12 months post-treatment entry
24 months post-treatment entry
30
24-month locked settings outcomes
100
90
90
129
80
70
75
72
60
days in locked
Mean number of
settings
50
57
42
40
30
20
22
20
10
0
GC
MTFC
Intervention group
12 months pre-treatment
12 months post-treatment
24 months post-treatment
31
Efficacy to Effectiveness and Implementation

• Project Keep

32
Project KEEPCascading Dissemination of a
Foster Parent Intervention

• A collaboration between the
• San Diego Health and Human Services Agency,
• Child and Adolescent Services Research Center,
• Oregon Social Learning Center, and
• funded by the National Institute of Mental
  Health.

33
The Goals of KEEP

• To increase the parenting skills of foster and
  kinship parents
• To decrease the number of placement disruptions
• To improve child outcomes
• To increase the number of positive placement
  changes (e.g. reunification, adoption)
• To test the cascade question-can the
  intervention be implemented with equal
  effectiveness by a second generation of
  interventionists?

34
Project KEEP aims to accomplish these goals by--

• Promoting the idea that foster parents can serve
  as key agents of change for children.
• Strengthening foster parents confidence and
  skill level so that they can successfully change
  their own and their childs behaviors.
• Helping foster parents use effective parent
  management strategies and provide them with
  support to do so
• Increasing short and long term positive child
  outcomes in multiple domains and settings
  home, school, with peers.

35
(No Transcript)
36
Support and Training for Foster Parents
37
Child Demographic Information
Mean
Age Age at Baseline 8.8 (SD2.2, n700)
Age Range 4-13 (n700)
Gender Female 52 (n364)
Male 48 (n336)
Ethnicity Caucasian 29 (n206)
African-American 25 (n175)
Hispanic/Latino/a 33 (n228)
Mixed/Other 13 (n91)
38
Foster Parent Demographic Information
Mean
Language Spoken English only 60 (n404)
Spanish only 8 (n52)
Both English and Spanish 32 (n122)
Household Income Less than 64,999 69 (n483)
Over 65,000 17 (n120)
Refused/Dont Know 14 (n97)
39
Foster Parent Demographics - 2
Mean
Employment Currently Employed (not including foster parenting) 49 (n343)
Employment of hours works per week (includes unemployed foster parents) 17.1 (SD20.8, n696)
Education Level High School/GED or less 41 (n285)
Education Level Some College 46 (n325)
Education Level Vocational or Technical Degree 1 (n48)
Education Level Bachelors Degree 7 (n48)
Education Level Graduate Degree 5 (n32)
40
Average Number of Children in Home
Control (n341) Treatment (n359) Both (n700)
Biological/Step Children .7 (1.1) .7 (1.2) .7 (1.2)
Adopted/Foster Children 2.4 (1.9) 2.6 (2.0) 2.5 (2.0)
Other Children .3 (1.1) .1 (.5) .2 (.9)
All Children 3.5 (2.0) 3.5 (1.8) 3.5 (1.9)
41
Rates of child problems Parent daily report
(PDR)

• PDR is collected by telephone from foster/kin
  parents
• Each call takes 5-10 minutes
• We collect 3 calls at baseline, another 3 calls 4
  months later, and a final 3 calls 6 months after
  that (10 months after baseline)
• PDR produces data on the occurrence of child
  behavior problems and foster parent stress

42
Greater than 5 problem behaviors per day at
baseline predicts placement disruption within the
next 6 months
After 5 behaviors, every additional behavior on
the PDR increases the probability of disruption
by 13
43
Good news We learned that

• Foster/kin parents tolerate about as much child
  problem behaviors as non-system families do 5
  behaviors
• PDR data is feasible to collect and is well
  tolerated by foster and kin parents
• PDR data tells you who to concentrate the
  intervention on given limited resources 40/60

44
(No Transcript)
45
PDR Total of Problem Behavior (All Children)
Baseline Termination
Control Group 5.8 (4.0) (n265) 5.4 (4.1) (n265)
Treatment Group 5.9 (4.3) (n 299) 4.4 (3.9) (n299)
P lt .05
46
PDR Total of Problem Behaviors(TC only)
Baseline Termination
Control Group 5.4 (3.7) (n229) 5.1 (3.9) (n229)
Treatment Group 5.7 (4.2) (n 241) 4.2 (3.8) (n241)
P lt .05
47
PDR Total of Problem Behaviors (Substitute
Child only)
Mean (SD)
Control Group 7.4 (4.9) (n36)
Treatment Group 5.1 (4.2) (n64)
48
PDR by interventionist cohort control
experimental
baseline term baseline term
developers N506 5.7 (4.0) 5.4 (4.2) 6.0 (4.3) 4.4 (4.0)
cascade N191 6.0 (3.9) 5.5 (4.1) 5.6 (4.2) 4.3 (3.6)
49
Intervention Effects on Parenting
Social Learning parenting practices Control BL T Treatment BL T
Uses rewards (1daily, 7never) 3.7 4.0 3.6 3.4
use point charts 18 22 26 51
who use time out 42 37 36 52
Doesnt warn/discuss 50 48 45 41
50
Conclusions

• Preliminary data suggests KEEP group
  participation increase foster and kin parent
  skills
• Increased parent skills translate into lower
  rates of child problem behaviors
• Lower rates of child problem behaviors translate
  into fewer placement disruptions more frequent
  placement transitions to family/relative care

51
Review of Design and Methods Issues Raised

• Use of Administrative Data to measure outcomes
  (CWS vs. Juvenile Court / School Records)
• Measurement of proximal or targeted processes
  (Parent Reports and PDR)
• Moving from Thick to Thin Assessment
• Power when large community units are unit of
  randomization
• How do we team up efficiently with PSMG